How Insurers Can Harness Artificial Intelligence What was once science fiction is fast becoming a fact of today’s business world. Artificial intelligence holds vast potential for insurers interested in reinventing their business models and transforming customer experience.

2 KEEP CHALLENGING July 2016

Executive Summary“The business plans of the next 10,000 startups are easy to forecast:

Take X and add AI.

This is a big deal, and now it’s here.”

– Kevin Kelly, founding executive editor of Wired magazine.1

Be it chat bots fueling customer service, Apple’s Siri answering quasi-existential

questions or Mark Zuckerberg trying to build a pocket-sized J.A.R.V.I.S, artificial

intelligence is touching our lives in many ways.

AI’s impact is poised to expand beyond these examples, as enterprise AI solutions

emerge to enhance operational efficiency, improve time-to-market, enable a

more intelligent way to sell and service, and more. During the last five years,

industrial use of AI – in terms of interest, investment, ideation and implementation

– has risen exponentially. Companies such as IBM, Apple, Toyota and Fidelity

have demonstrated interest and appetite for deep research and innovation by

introducing AI platforms and solutions for customers, partners and employees. For

instance, Toyota’s websites use AI to perform sophisticated, real-time reasoning to

ensure manufacturing feasibility or inventory availability for the exact combination

of options chosen by a consumer to design a specific vehicle.2

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 3

At the same time, front runners across industries are partnering with technology

companies to identify game-changing business solutions that can be achieved

through AI. An example is The North Face, which is experimenting with its Fluid

Expert Personal Shopper tool, powered by IBM’s Watson, to provide customers

with a more intuitive search experience through a natural language capability.3

AI applications are continuously evolving, and numerous success stories are

emerging across geographies and lines of business. (For more on this topic, see our

Cognizanti article “Where We Stand and Where We’re Going with Automation.”4)

The insurance industry has not been immune to AI’s advancement – whether

implementing robo-advisors for investment management (i.e., Vanguard5 and

Charles Schwab6) or applying AI to insurance and loan underwriting (i.e., the

Chinese search giant Baidu,7 which provides enhanced risk assessment capabilities).

AI, however, also comes with its own set of caveats that insurers must understand

and address before initiating a program. This white paper focuses on potential

insurance applications of AI and the benefits that carriers can realize by

implementing AI in their processes. We also evaluate AI challenges, such as its

never-ending “learning loop,” the huge upfront resource investment it requires and

its adoption and reliability issues, to name a few.

4 KEEP CHALLENGING July 2016

The Evolution of AI At its essence, AI is about embedding human intelligence into machines, enabling systems to learn, adapt and develop solutions to problems on their own.

Various AI-related technologies, such as natural language processing (NLP), computer vision, robotics, machine learning and speech recognition, have substan-tially progressed over the years to coalesce into systems that do, think, learn and continuously adapt (see Figure 1).

Just five years ago, leading experts at MIT were confident that cars would never be able to drive themselves.8 Today, Uber9 has successfully picked up the torch ignited by Google’s autonomous automobile initiatives, and Apple is not too far away.10 This is just one example of how quickly things can advance when researchers and investors alike envision the cross-industry implications of a disruptive force. In fact, significant commercial activity is underway in AI that is affecting organizations in every sector – or will soon.

Private investments in AI have expanded an average of 62% annually since 2011.11 Banking and financial services, healthcare and the travel industry are leading the

Figure 1

Categorizing AI

-

• Characterized by traditional software, robotic arms and process automation tools.

• Key theme: Automation

• Enablement of high-volume, rules-based activities.

• Characterized by the need to identify anomalies, uncoded scenarios and critical incidents. • Key themes: Intensive research and experimentation in AI.

• Enablement of man-machine learning.

• Characterized by self learning, highly dynamic, non-rules-based adapting systems.

• Key themes: Improving accuracy, piloting systems and the evolution of commercial AI systems.

• Simple rules-based processes were the primary areas of focus in the insurance industry.

• Rules-based automated underwriting, lead generation, rules-based advisory services.

• Rapid adoption of technology in the form of IoT, Code Halos, big data, social listening.

• Foundational stage for insurers to enable systems to better understand customers and what's happening in the market and develop into systems that think.

• Use of AI to aid human workers.

• Development of the ecosystem for AI.

• Evolution from man-machine learning to dynamic underwriting, virtual assistants, robo-advisory and other viable use cases of AI.

SYSTEMS THAT DO SYSTEMS THAT THINK

TIME

SYSTEMS THAT LEARN

2016

ARTIFICIALINTELLIGENCE

INDUSTRY

INSURANCEINDUSTRY

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 5

way with AI platform pilots to identify solutions to business problems. In December 2014, total assets under management with robo-advisors in the U.S. was $19 billion, and is estimated to reach $2 trillion by 2020.12

In addition to Uber and Apple, others such as Lyft,14 Didi Chuxing15 and several automobile giants16 have all made investments in devising driverless cars. In healthcare, organizations are using AI solutions to enhance diagnoses, real-time patient monitoring17 and adherence to medication treatments. In fact, some companies are building virtual assistants for doctors.18 Moreover, retail giants are applying AI-fueled customer insights to design new fashion lines.19

With so much activity around AI experimentation and implementation – combined with customer demand, cost pressure and the need to maintain or expand their foothold in the market – insurers can no longer afford to overlook AI and its potential implications.

Based on our research and analysis, we recommend that insurers approach AI adoption in two stages.

• Stage 1: Use AI to assist human workers rather than displacing them, particularly in two areas:

> Underwriting. AI systems can be used to perform research, aggregate, refine and present required information to underwriters, allowing them to focus on core underwriting activities.

> Advisory services. Virtual assistants can manage the low-value activities of advisors, such as lead management, scheduling, planning, licensing, etc., enabling them to focus on building skills and providing value-added services.

While insurers reap the benefits of man-machine learning, they also need to make a parallel effort to connect the various technology components and stakeholders across the insurance value chain and develop an ecosystem for a full-fledged launch of AI services. An example would be devising a strategy to capture unstructured information from social media, blogs, customer interac-tions, external data agencies, internal processes, etc. The key focus for insurers at this stage is to build up a central repository that acts as a knowledge base to be exposed to AI systems for rigorous training and tuning.

• Stage 2: With AI making inroads in the insurance industry and the peripheral systems elevated to support AI, insurers can evaluate pilot programs that aim to turn underwriting claims into dynamic self-learning models. Transforming the customer experience through virtual assistants, robo-advisory, robo-contact centers and chat bots can be explored in the next five to 10 years.

Insurance Industry Implications: From ‘So What,’ to ‘Now What’ AI’s potential spans all insurance areas. Continuous insights that emerge through the perpetual churning of structured and unstructured information (as shown in Figure 2, next page) can elevate an organization’s ability to better understand changing market dynamics, competitor activities and, most importantly, customer wants, needs and desires with unprecedented granularity.

Accrued benefits include:

• Revenue expansion: The real-time learning and adaptive capabilities of AI provide a ready platform for insurers to explore new product lines, geographies and cus-tomer segments, as well as quickly identify new avenues for revenue expansion.

Organizations can scale exponentially by increasing their ability to manage areas such as information search, data management and process automation, delivered by virtual assistants across business processes.

6 KEEP CHALLENGING July 2016

• Advisory excellence: Robo-advisors hold the potential to dramatically change the dynamics of insurance advisory by not only eliminating the drawbacks of human advisors but also assisting them to develop and hone their skills. Robo- advisors eliminate human bias and improve trust and confidence levels by pro-viding consistent, rules-based advisory services at an affordable cost. To improve their job performance, successful human advisors will use virtual assistants to tackle routine tasks and focus on more constructive, creative and socially inter-active challenges. Doing so will enable them to exhibit 24x7 availability to the external world.

• Improved operational efficiency: AI-based systems offer a host of opportunities for the insurance industry to improve its operational efficiency. This includes:

> Reduced turnaround time: AI-based needs analysis systems allow insurers to not only improve their probability of lead-to-quote conversion but also re-duce turnaround time (TAT) conversions.

> Lower costs: AI solutions enable organizations to reduce their manpower requirements and thus benefit from significant savings in overhead costs, especially those associated with routine jobs.

> Improved productivity: With AI systems performing routine activities, em-ployees can focus on skilled tasks, building expertise and evolving the AI solu-tions. Cumulatively, insurers can expect a surge in operational efficiency with AI solutions while realizing non-tangible benefits in terms of:

» Faster access to information.

» Elimination of subjectivity in response or actions from employees.

» Reduced need for system alterations with the ever changing environment.

• Maximized customer experience: With NLP, speech recognition and virtual as-sistants (robos), insurers can embrace innovative ways of transforming the cus-

Figure 2

The Anatomy of an AI Solution

SUPPORT SYSTEMS

KnowledgeDatabase

SMAC

Applications

DATA SOURCES

Data AgenciesApplications Regulators

Social Media Public DomainUnstructuredData

Third-PartyEmployees

InternalUsers

Advisors Customers

Artificial Intelligence

ManualIntervention

MachineLearning

VirtualAssistants

Speech/textConversion

FaceRecognition

Natural LanguageProcessing

END USERS

DATA INGESTION

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 7

tomer experience. With virtual assistants available at various touchpoints, insur-ers can take their customer service to greater heights by offering contextual and personalized products and solutions, accomplishing “first-time right” resolutions for queries and complaints, and providing timely nudges and reminders to com-plete necessary transactions to maintain financial wellness.

• Competitive advantage: Insurers with AI capabilities can position themselves to handle market challenges better than their competitors in the ever-changing insurance business. By leveraging machine learning and the ability to read un-structured data, insurers can develop a real-time appreciation of prospects’ be-havioral and demographic actions, recognize imperceptible changes in the mar-ket forces that dictate those changes and forecast optimal responses.

Viewing AI through Multiple Lenses Any complex, evolving technology that requires extensive decision-making, emotional connections and visible, long-term consequences is intrinsically tied to the lean startup method proposed by Eric Ries in his similarly titled book.20 This model is built on a continuous loop of feedback and learning that feeds into an engine of continuous development.

This model applies particularly well to insurance, as it is a “human-centric,” push-based, emotion-driven industry. Business goals such as operational efficiency and revenue expansion must go hand-in-hand with best-in-class customer experience for predicting, preempting and indemnifying (in principle) intensive customer needs. While some companies have made initial steps with chat bots, many have also advanced to robo-advisory. Based on organizational maturity and market appetite, insurers can find themselves in one of three AI stages: foundational, incre-mental and institutional.

The Foundationally Intelligent Insurer Insurers at this stage rely heavily on traditional processes and legacy systems. Most have a limited online and media social presence and tend to resist major organiza-tional change. As a result, the major imperatives for these insurers to gain a solid foothold in the market include:

• Attaining rapid growth through channel and touchpoint expansion.

• Employing operational intelligence to optimize resource overhead and improve customer experience.

• Modernizing technology to solve legacy problems such as scalability, service turnaround times, etc.

> Our take: Conventional wisdom dictates that these insurers should not invest in implementing something as intricate and resource-intensive as AI and in-stead concentrate on more elementary techno-business issues that promise a better return on investment. However, closer inspection may reveal that the insurers’ imperatives can be satisfied to some extent through easily-configu-rable and implementable solutions using AI, such as:

Insurers with AI capabilities can position themselves to handle market challenges better than their competitors in the ever-changing insurance business.

8 KEEP CHALLENGING July 2016

» Chat bots (starting with a “smart” configuration rather than NLP-enabled).

» Personal financial trackers and advisors (starting with a rules-based con-figuration) to target prospects and customers.

Thus, through the use of rules-enabled, easily configurable decision and process engines, insurers can not only save on resources but also ensure more efficient marketing, lead management and sales efforts – all at minimal cost to the company. These efforts will also pave the way for the next wave of sentience.

The Incrementally Intelligent Insurer Insurers at this stage have embarked on the digital journey and have improved engagement with distribution partners, customers and internal stakeholders. Such insurers have invested heavily in IT solutions that enable “pull” marketing techniques rather than just “push” approaches. Targeted marketing, gamification, a social media presence, mobile solutions and analytics are the key themes within the organization at this stage. Insurers in this category are typically characterized by:

• Knowledge of customers’ needs and behaviors.

• Multiple channels and touchpoints of sales and service, which introduce new challenges for enabling the customer experience.

> Our take: Improvise on current business processes by building a strong foun-dation to be ready for AI takeoff. The key driver for these insurers is market consolidation at minimal cost. Examples of basic AI pilot solutions that can be built and tested in a short time span with a relatively low risk of investment and failure include:

» Virtual sales assistants that manage basic routine work (e-mails, meetings, lead search, etc.).

» Automated algorithms for needs analysis that can be deployed across all customer-facing channels, thus ushering in robo-advisors.

The Institutionally Intelligent Insurer Insurers at this stage are at the forefront of the insurance industry, employing technology to effectively solve business problems. These businesses have advanced point-of-sale capabilities, straight-through processing (STP) functionality and a single view of the customer across all channels, among other characteristics. The most important techno-business levers of insurers at this level include:

• Improving the customer experience at every touchpoint and channel.

• Discovering possible routes of up- and cross-sell through better customer data and interaction mining.

• Improving operations through intelligent decisions across manually intensive processes such as underwriting, claims management, etc.

> Our take: This is the playground for insurers looking to disrupt business mod-els, processes and products. These businesses should think big, start small, fail fast and scale rapidly to achieve appreciable market dominance. They have developed the foundation and are equipped to explore complex AI busi-ness solutions, such as:

» A dynamic underwriting model based on machine learning to provide enhanced context relevancy for decision-making.

» Claims transformation through intelligent prediction and adjudication.

» Contact center modernization through voice recognition and interactions mining.

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 9

Insurers on the BandwagonOur experience with insurance companies reveals a strong desire to employ AI to solve operational challenges, whether through in-house capability development; acquisition of a boutique AI product firm that specializes in a specific set of capa-bilities; or partnership with vendors to co-invest in building required capabilities and platforms.

Some insurers have already adopted AI capabilities through one of these methods. For example:

• Swiss Reinsurance Co. is working with IBM Watson to develop a range of un-derwriting solutions and achieve accurate risk pricing.

21

• Insurify.com enables customers to generate price quotes by texting a photo of their license plate. Insurify’s patent-pending license plate option Evia (short for expert virtual insurance agent) is built on machine learning and natural language processing. It does not require a mobile app to be downloaded or installed; consumers can simply text a photo of their license plate.

22

• Manulife has transformed its authentication mecha-nism from the conventional password and PIN system to voice recognition. Through a partnership with voice recognition leader Nuance Communications, Manulife analyzes unique voice characteristics to create individual “voiceprints” for customers. When customers call in to ac-cess their accounts, their voice is compared with a stored voiceprint. If there’s a match, access is granted.

23

• Customers at Buyonic Insurance Agency in Austin, TX, are greeted not by a human but by Siber, the buff, blue-eyed, bald company principal. Siber can rate, bind and issue policies on the spot, while answering phones and making robocalls.

24

• ZestFinance takes an entirely different approach to underwriting by combin-ing machine-learning techniques and data analysis with traditional credit scor-ing.

25

• Microsoft and GAFFEY Healthcare have collaborated on a pilot to deploy machine learning in GAFFEY’s claims automation and processing engine at hos-pitals.

26

• USAA added Nuance Communications’ virtual assistant (Nina) to its exist-ing mobile customer service apps to enable speech recognition, text-to-speech, voice biometrics and NPL capabilities.

27

Syncing with AI’s Inexorable RiseTo ride the AI wave, insurers need to equip themselves with the ability to gather specific information from various sources. They also need a highly integrated and digitized environment, as well as a dynamic workforce ready to take on innovation head on. Our current stable of products and platforms, including Life EngageTM (see Figure 3, next page), Geolocus, Social Prism and Interactions Insights Hub, can be equipped with AI capabilities to provide insurer-ready components across functional areas such as new business, sales and distribution, and customer service.

Our experience with insurers reveals a strong desire to employ AI to solve operational challenges, whether through in-house capability development; acquisition of a boutique AI product firm; or partnership with vendors.

10 KEEP CHALLENGING July 2016

Interactions Insight Hub A few years ago, the over-riding goal of the insurance industry was to derive key insights from each and every customer touchpoint. Our Interactions Insight Hub is a platform built on the concept of Code HaloTM 28 thinking. It mines customers’ interactions with insurers and combines these with other identifiers derived from internal and externally available data to derive actionable insights. The insights are used to attract and retain customers, as well as upsell and cross-sell.

The solution has been piloted by a few insurers, including one that improved its retention of premium customers and another that boosted its customer satis-faction scores. The existing platform can be further strengthened with machine learning, speech-to-text conversion and sentiment analysis capabilities to serve myriad customer-serving purposes.

Our Interaction Insight Hub acts as a rich data warehouse, comprising customer conversations across numerous transactions and contexts for insurers, creating a valuable knowledge base that can be used to train and develop any AI solution, especially customer-facing solutions, such as virtual assistants, contact center, robo-advisor, etc.

Partnering with clients, we are exploring various business areas for which automated actions can be triggered based on Code Halo-derived insights. One example would be to notify contact center agents, robo-advisors, virtual assistants, human advisors and other stakeholders across the organization in real time when the system uses

Figure 3

• Explore avenues of information sources to gather lifestyle, professional and life event-based details for a better understanding of risk profiles.

• Shift from a customer segment approach to an individual customer model to understand unique needs of each customer.

• Use facial recognition to identify customers and gather associated details.

Cognizant Life Engage is a unified life insurance point-of-sale, underwriting and customer service platform that engages customers with an efficient, modern and seamless sales experience.

Benefits> Customer delight.

> Consultative selling.

> Improved conversion ratio.

> Speed to market for new products.

> Pricing accuracy.

> Improved customer trust.

> Personalized and consistent advisory services.

> Ability for advisors to focus on developing advisory skills.

> 24x7 availability of the sales force for customers through virtual assistants.

> Improved digital presence for advisors.

• Use virtual assistants for sales self-service.

• Use NLP to serve notifications and seek responses and next steps of action.

• Use virtual assistants for agents to identify leads, interact over e-mail and schedule meetings.

• Use NLP to interact with customers for needs analysis.

• Use contextual and reflexive questions based on customer interest and response during needs analysis.

• Identify a host of variables to be fed into a dynamic, non-linear underwriting model.

AI Capability Area

Cognizant Life Engage

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 11

metadata and other information to identify a high-net-worth customer who is unhappy about a product, service or transaction The immediate learning expected at each touchpoint would be to identify and resolve the customer’s cause of incon-venience and take restorative measures on behalf of the organization.

Social PrismOur Social Prism accelerator is based on the concept of using social listening to uncover critical insights that can inform and shape an insurer’s sales and service strategy.29

Insurers can manage their brands effectively on social media using machine learning in conjunction with Social Prism. Based on customer sentiment identified through Social Prism, AI mechanisms can respond appropriately in real time. When the comments involve a service issue, an AI mechanism can refer the customer to a corresponding service team. The insights gathered from Social Prism can act as potential knowledge feed for various business processes, such as brand management, product management and trend analysis. Robo-advisors or virtual assistants, when coupled with Social Prism, can be used as a real-time platform for advisors to stay connected with customers.

Challenges on the Path Less Traveled Given the extraordinary possibilities of AI, it’s no wonder that many insurers are jumping on the bandwagon. However, hype-driven, ill-informed investments can lead to loss and disappointment, while appropriate invest-ments can dramatically improve performance and create competitive advantage. The following sections provide a critical look at AI’s clear and present dangers that could undermine planning and implementation, as well as recommendations to help insurers achieve their objectives.

Challenge #1: Building a Strong FoundationIt took years for IBM and IPsoft to build AI platforms, and innovation leaders are finally piloting these platforms to assess their potential benefits. IBM has committed $1 billion to commercialize Watson, and IPsoft has spent almost 16 years developing and nurturing AI.30, 31

However, for AI platforms to solve business problems, they need to be exposed to huge volumes of domain-specific information covering all possible business scenarios. The success of an AI solution largely depends on the continuous learning it creates from every single business transaction or interaction it makes. The challenge for an insurer is to ensure the availability and accuracy of information to be fed into the AI-based solutions.

Most insurers will seek a cloud solution, provided by a technology company, that needs to be familiarized and trained with insurer-specific information containing product, process, transaction and application information. Insurers need to ensure they have the appropriate training data set – large enough to learn from and varied enough to train from.

For example, if an AI solution is aimed at replacing the contact center, the bots need to be fed a wide variety of customer query and response data that informs expected workflows at the contact center. The insurance industry, in collaboration

For AI platforms to solve business problems, they need to be exposed to huge volumes of domain-specific information covering all possible business scenarios.

12 KEEP CHALLENGING July 2016

with technology and consulting partners, will need to build an ever-evolving com-prehensive knowledge repository that would not only aim to collate domain infor-mation specific to each business function poised to leverage AI in the near future, but also serve as a treasure trove of customers’ changing demographics and psy-chographics, as well as how these changes affect their interactions with insurers.

Challenge #2: The Glitch Risk

• Reliability: Microsoft’s “teen girl’ AI chat bot32 quickly became infamous for its use of deeply racist and sexist tweets. Facebook claims it can recognize faces with 97% accuracy — but it still cannot generally recognize multiple objects in a scene or reliably understand the actions it is witnessing.33 Still in the pilot phase with many organizations, advanced AI needs to battle the challenges of accuracy and getting things right the first time.

Technologies such as speech recognition and machine learning require human oversight for their work to equate with human capabilities. Voice rec-ognition systems require painstaking training and could only work well with controlled vocabularies. The success of NLP and speech recognition depends

on their accuracy and ability to cope with diverse accents, background noise and distinctions between homophones (such as “buy” and “by”), as well as the need to work at the speed of natural speech.

AI systems introduce the risk of unpredictable and bizarre errors that are not easily resolved through root cause analysis. When a machine is incorrect, it can be wrong in a far more dramatic way, with more unpredictable outcomes, than a human could. Not only may AI systems produce imperfect results, but they may also require a significant investment of human time to re-train or re-configure before resuming their work.

• Contextualization: Another important aspect of any AI system that interacts with people is its need to understand people’s intentions rather than carrying out commands lit-erally. An AI system must analyze and understand whether

the behavior a human is requesting is “normal” or “reasonable.” Suppose a self-driving car is told to “get us to the airport as quickly as possible.” Would the autonomous driving system increase its speed to 125 mph, putting pedestrians and other drivers at risk? In addition to relying on internal mechanisms to ensure proper behavior, AI systems need the ability – and responsibility – of working with people to obtain feedback and guidance. They must know when to stop and “ask for directions” and always be open to feedback.

In their initial phase, AI solutions will require insurers to make provisions for regular manual interventions, thereby spending a lot of time, energy and money on monitoring AI and rectifying deviations. Staffing teams with the talent required to pilot and nurture such systems will be a big challenge.

Not only may AI systems produce

imperfect results, but they may also require a

significant investment of human time to re-train or re-configure before

resuming their work.

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 13

Challenge #3: Stakeholder Readiness

• Workforce: Insurers willing to implement AI solutions will need to carefully manage the threat to their workforce by redesigning tasks, jobs, management practices and performance goals. From an insurer standpoint, advisory, opera-tions and contact center teams will undergo a major transition to align with AI implementations. Employees in these departments will need to build expertise and evaluate alternative work profiles.

• Customers: Organizations will need to evaluate the customer mindset toward and readiness for AI. Insur-ers can do this by asking themselves a few questions that place them in the customer’s mindset, such as:

> Can I trust a machine in making a long-term finan-cial decision?

> Will bots and AI be capable of dealing with my emotions, especially during complaints and claims?

> How reliable is the AI system vis-à-vis the traditional systems and processes that used to serve me well?

Challenge #4: Privacy Concerns/Regulatory HurdlesSince AI solutions require every interaction/transaction to be recorded for machine learning, the insurance industry will have to battle data privacy concerns since most AI solutions are likely to reside on the cloud of a third-party technology provider. Insurers will need to ensure that information security and data privacy policies, procedures, methods and tools are employed to protect data from breach or unintended use.

Keeping a close watch on the regulatory changes will be a big challenge for insurers, considering the lag of regulatory bodies in embracing technological innovations. Complete know-how of the material legal risks associated with new AI technology is a major concern for the nascent AI-enabled insurance industry.

Challenge #5: Technology RefreshImplementing AI will require a good deal of supporting technology and maturity of the insurer technology landscape. Among the attributes insurers will need to evaluate are the existing state of the technology landscape, the level of integration between various systems and applications, the state of process digitization and availability of data from various sources.

Building such infrastructure and enabling integration with AI solutions will require both time and cost considerations.

Looking AheadBillions of dollars have been invested to pilot and commercialize the technologies that drive the gamut of AI capabilities; as is evidenced by the examples above, insurers are beginning to realize benefits. However, while it might appear that AI is a cure-all for many sales, service and risk management conundrums, insurers should follow a series of self-diagnostic steps before embarking on this journey.

• Assess readiness: This can be done by socializing the concept and experience of AI solutions and soliciting response though surveys, interviews and discussions. At the same time, AI decision-makers must spend quality time with their execu-

Since AI solutions require every interaction/transaction to be recorded for machine learning, the insurance industry will have to battle data privacy concerns.

14 KEEP CHALLENGING July 2016

tive teams, peers and functional business leaders to reflect upon the potential implications of new AI-based products or applications on operating models, products and operational workflow. Moreover, decision-makers must also assess the Internal technology landscape and the extent of historical data available in terms of quantity and quality, structured and unstructured data (from internal and external sources), as well as the ways and means to scale and support AI applications.

• Start small: Our experience and research suggest that given the cultural and risk challenges facing the sector, insurers should start by developing a proof of concept model that can safely be tested and adapted in a risk-free environment. Since AI machines excel at routine tasks and their algorithms often learn over

time, insurers should focus their early efforts on the pro-cesses or assessments that are widely understood and add a modicum of value. With the knowledge gathered from the assessment phase, insurers should then identify the use case for a proof of concept, considering the activities that require agility, automation and continuous innovation. A second step would be to identify the right technology part-ner and AI solution to transform the identified use case from concept to reality.

• Manage change: Because AI capabilities can poten-tially displace humans (or require talent upskilling), in-surers need an effective and thoughtful HR strategy. Full communication and retraining of affected staff, as well as a focus on building new skill sets and training, will go a long way toward minimizing resistance and encouraging acceptance. This means insurers must focus on effective

change management to ensure that impacted employees understand the tools are deployed to help them do a better job, increase their productivity and value, and enhance customer satisfaction, which in turn will raise employee satisfac-tion and retention.

Insurers and their employees must collectively believe in the potential ampli-fication of productivity possible through human-augmented AI solutions. The success of an AI solution depends on continuous learning and tuning of the AI model; hence, the concept of learn-unlearn-relearn will play a special role at this stage. The more decisions machines make, and the more data they analyze, the better equipped they are to undertake increasingly complex tasks and deliver more accurate and appropriate actions.

Since AI machines excel at routine tasks and their

algorithms often learn over time, insurers should focus

their early efforts on the processes or assessments

that are widely understood and add a modicum of value.

Note: Code HaloTM is a trademark of Cognizant Technology Solutions. All company names, trade names, trademarks, trade dress, designs/logos, copyrights, images and products referenced in this white paper are the property of their respective owners. No company referenced in this white paper sponsored this white paper or the contents thereof.

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 15

Footnotes1 Kevin Kelly, “The Three Breakthroughs that Have Finally Unleashed AI on the World, ” Wired, Oct. 27,

2014, http://www.wired.com/2014/10/future-of-artificial-intelligence/.

2 Matthew Chin, “Artificial Intelligence Framework Now Powers Toyota Websites,” Phys.org, May 13, 2016, http://phys.org/news/2016-05-artificial-intelligence-framework-powers-toyota.html.

3 Rachel Arthur, “Future of Retail: Artificial Intelligence and Virtual Reality Have Big Roles To Play,” Forbes, Jan 15, 2016, http://www.forbes.com/sites/rachelarthur/2016/06/15/future-of-retail-artificial-intelligence-and-virtual-reality-have-big-roles-to-play/#7357b03f420c.

4 Matthew Smith, “Intelligent Automation: Where We Stand — and Where We’re Going,” Cognizanti, Vol. 9, Issue 1, 2016, https://www.cognizant.com/content/dam/Cognizant_Dotcom/whitepapers/bots-at-the-gate-intelligent-automation-where-we-stand-and-where-we-are-going-cognizanti12-codex2096.pdf.

5 Margaret Collins, “The Vanguard Cyborg Takeover,” Bloomberg, March 24, 2016, http://www.bloomberg.com/news/articles/2016-03-24/the-vanguard-cyborg-takeover.

6 Trevor Hunnicutt, “Schwab ‘Robo Adviser’ Grows to $5.3 Billion in its Debut Year,” Reuters, Jan. 6, 2016, http://www.reuters.com/article/us-charles-schwab-automation-idUSKBN0UK2GA20160106.

7 Jeremy Kahn, “Baidu Looks to Artificial Intelligence to Reduce Insurance Risks,” Bloomberg, Jan. 20, 2016, http://www.bloomberg.com/news/articles/2016-01-20/baidu-looks-to-artificial-intelligence-to-reduce-insurance-risks.

8 Will Knight, “Driverless Cars Are Further Away than You Think,” MIT Technology Review, Oct. 22, 2013, https://www.technologyreview.com/s/520431/driverless-cars-are-further-away-than-you-think/.

9 “Steel City’s New Wheels,” Uber website, May 19, 2016, https://newsroom.uber.com/us-pennsylvania/new-wheels/.

10 Daisuke Wakabayashi and Douglas MacMillan, “Apple’s Latest $1 Billion Bet Is on the Future of Cars,” The Wall Street Journal, May 14, 2016, http://www.wsj.com/articles/apples-1-billion-didi-investment-revs-up-autonomous-car-push-1463154162.

11 “The Future of Artificial Intelligence,” Trustmarque, Dec. 8, 2015, http://www.trustmarque.com/future-artificial-intelligence/.

12 Michael Regan, “Robo Advisers to Run $2 Trillion by 2020 if this Model Is Right,” Bloomberg, June 18, 2015, http://www.bloomberg.com/news/articles/2015-06-18/robo-advisers-to-run-2-trillion-by-2020-if-this-model-is-right.

13 Rob Davies, “Lyft to Trial Driverless Cars on U.S. Roads,” The Guardian, May 6, 2016, https://www.the-guardian.com/business/2016/may/06/lyft-driverless-cars-uber-taxis-us-roads-chevrolet-bolt.

14 Daisuke Wakabayashi and Douglas MacMillan, “Apple’s Latest $1 Billion Bet Is on the Future of Cars,” The Wall Street Journal, May 14, 2016, http://www.wsj.com/articles/apples-1-billion-didi-investment-revs-up-autonomous-car-push-1463154162.

15 Bill Vlasic, “Ford and Google Team Up to Support Driverless Cars,” The New York Times, April 27, 2016, http://www.nytimes.com/2016/04/28/business/ford-and-google-team-up-tosupport-driverless-cars.html?_r=0.

16 Daniel B. Neill, “Using Artificial Intelligence to Improve Hospital Inpatient Care,” Carnegie Mellon University, 2013, http://www.cs.cmu.edu/~neill/papers/ieee-is2013.pdf.

17 Daniela Hernandez, “Artificial Intelligence Is Now Telling Doctors How To Treat You,” Wired, June 6, 2014, http://www.wired.com/2014/06/ai-healthcare/.

18 Samidha Sharma, “Artificial intelligence to Design Myntra’s New Fashion Line,” The Times of India, Aug. 15, 2015, http://timesofindia.indiatimes.com/tech/tech-news/Artificial-intelligence-to-design-Myntras-new-fashion-line/articleshow/48489014.cms.

19 Eric Ries, The Lean Startup, Crown Business, Sept. 13, 2011, http://theleanstartup.com/book.

16 KEEP CHALLENGING July 2016

20 “Swiss Re Uses Cognitive Computing to Improve its Underwriting Process,” IBM, Feb. 16, 2016, http://www.ibmbigdatahub.com/video/swiss-re-uses-cognitive-com-puting-improve-its-underwriting-process.

21 Andrew G. Simpson, “Virtual Insurance Agent Insurify with License Plate Quote Feature Raises $2 Million,” Insurance Journal, Jan. 28, 2016, http://www.insurance-journal.com/news/national/2016/01/28/396709.htm.

22 Susan Yellin, “Manulife Customers Can Say ‘So Long’ or ’Au Revoir’ to Passwords and PINs,” Insurance Journal, Oct. 5, 2015, http://insurance-journal.ca/article/voice-recognition-system-allows-manulifes-clients-to-skip-pins-and-passwords/.

23 Lirpa Loof, “First Robot-Run Insurance Agency Opens for Business,” Insurance Journal, April 1, 2016, http://www.insurancejournal.com/news/national/2016/04/01/403804.htm.

24 Leena Rao, “ZestFinance Debuts New Data Underwriting Model to Ensure Lower Consumer Loan Default Rates,” Tech Crunch, Nov. 19, 2012, http://techcrunch.com/2012/11/19/zestfinance-debuts-new-data-underwriting-model-to-ensure-lower-consumer-loan-default-rates/.

25 “GAFFEY Healthcare Announces Utilization of Microsoft’s Machine Learning Predictive Analytics to Improve Collections Automation Platform – GAFFEY AlphaCollector,” Street Insider, May 6, 2015, http://www.streetinsider.com/Press+Releases/GAFFEY+Healthcare+Announces+Utilization+of+Microsofts+Machine+Learning+Predictive+Analytics+to+Improve+Collections+Automation+Platform+%E2%80%93+GAFFEY+AlphaCollector/10531436.html.

26 “USAA and Nuance Create Financial Literacy Virtual Assistant to Help Millennials Break Negative Savings Rate,” Business Wire, July 22, 2015, http://www.business-wire.com/news/home/20150722005163/en/USAA-Nuance-Create-Financial-Litera-cy-Virtual-Assistant

27 Malcolm Frank, Paul Roehrig and Ben Pring, Code Halos: How the Digital Lives of People, Things, and Organizations Are Changing the Rules of Business, John Wiley & Sons. April 2014, http://www.wiley.com/WileyCDA/WileyTitle/productCd- 1118862074.html.

28 Social listening is the process of monitoring digital media channels to devise a strategy that will better influence consumers.

29 “IBM Watson Group Unveils Cloud-Delivered Watson Services to Transform Industrial R&D, Visualize Big Data Insights and Fuel Analytics Exploration,” IBM Press Release, Jan. 9, 2014, http://www-03.ibm.com/press/us/en/pressre-lease/42869.wss.

30 “Accenture and IPsoft Launch Accenture Amelia Practice to Help Organizations Accelerate Adoption of Artificial Intelligence,” IPsoft press release, May 16, 2016, http://www.ipsoft.com/2016/05/16/accenture-and-ipsoft-launch-accenture-amelia-practice-to-help-organizations-accelerate-adoption-of-artificial-intelligence/.

31 Helena Horton, “Microsoft Deletes ‘Teen Girl’ AI After It Became a Hitler-loving Sex Robot within 24 Hours,” The Telegraph, March 24, 2016, http://www.telegraph.co.uk/technology/2016/03/24/microsofts-teen-girl-ai-turns-into-a-hitler-loving-sex-robot-wit/.

32 Amit Chowdhry, “Facebook’s DeepFace Software Can Match Faces With 97.25% Accuracy,” Forbes, March 18, 2014, http://www.forbes.com/sites/amitchow-dhry/2014/03/18/facebooks-deepface-software-can-match-faces-with-97-25-ac-curacy/#6c2bda1127e5.

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 17

About the AuthorsSrinivasan Somasundaram is a Director within Cognizant Business Consult-ing‘s Insurance Practice. He has 20-plus years of experience in the life, annuity and pension sectors. Srini’s experience includes managing insurance operations, business consulting and product development, business transformation program delivery, developing business propositions, GTM and large deal support. He has worked on projects in the UK, U.S., South Africa and APAC, with a focus on sales and distribution. Srini holds a post-graduate degree in management from Bhara-thithasan University. He can be reached at Srinivasan.Somasundaram@cognizant.com | http://in.linkedin.com/pub/srinivasan-somasundaram/0/a5a/156.

Aritro Bhattacharya is a Manager within Cognizant Business Consulting‘s Insurance Practice. He has nearly 10 years of experience in the life, annuity and pension sectors, including consulting and delivery management, insurance IT product development, organizational change management, functional implementation and migration. He also has experience in social media consulting, digital distribution transforma-tion, digital marketing and sales consulting and digital innovation. Aritro holds a post-graduate degree in management from the Indian School of Business. He can be reached at Aritro.Bhattacharya@cognizant.com | https://www.linkedin.com/pub/aritro-bhattacharya/21/34b/629.

Divyaprakash Modi is a Consultant within Cognizant Business Consulting‘s Insurance Practice. He has six years of experience in the insurance industry across functions such as life insurance policy servicing, group business, life insurance IT consulting and business analysis. As part of his exposure to the global insurance industry, Divy-aprakash has worked with global insurers and insurers based in India. He holds a post-graduate diploma in business management from Sydenham Institute of Management Studies. He can be reached at Divyaprakash.Modi@cognizant.com | https://in.linkedin.com/pub/divyaprakash-modi/39/357/891.

18 KEEP CHALLENGING July 2016

HOW INSURERS CAN HARNESS ARTIFICIAL INTELLIGENCE 19

World Headquarters500 Frank W. Burr Blvd.Teaneck, NJ 07666 USAPhone: +1 201 801 0233

Fax: +1 201 801 0243Toll Free: +1 888 937 3277

inquiry@cognizant.com

European Headquarters1 Kingdom Street

Paddington CentralLondon W2 6BD

Phone: +44 (0) 207 297 7600Fax: +44 (0) 207 121 0102

infouk@cognizant.com

India Operations Headquarters#5/535, Old Mahabalipuram Road

Okkiyam Pettai, ThoraipakkamChennai, 600 096 India

Phone: +91 (0) 44 4209 6000Fax: +91 (0) 44 4209 6060

inquiryindia@cognizant.com

© Copyright 2016, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the express written permission from Cognizant. The information contained herein is subject to change without notice. All other trademarks mentioned herein are the property of their respective owners.

Codex 2131

About Cognizant Business ConsultingWith over 5,500 consultants worldwide, Cognizant Business Consulting offers high-value digital business and IT consulting services that improve business performance and operational productivity while lowering operational costs. Clients leverage our deep industry experience, strategy and transformation ca-pabilities, and analytical insights to help improve productivity, drive business transformation and increase shareholder value across the enterprise. To learn more, please visit www.cognizant. com/consulting or e-mail us at inquiry@cognizant.com.

About Cognizant’s Insurance Business UnitCognizant is a leading global services partner for the insurance industry. In fact, seven of the top 10 global insurers and 33 of the top 50 U.S. insurers benefit from our integrated services portfo-lio. We help our clients run better by driving greater efficiency and effectiveness, while simultaneously helping them run differ-ently by innovating and transforming their businesses for the future. Cognizant redefines the way its clients operate — from increasing sales and marketing effectiveness, to driving process improvements and modernizing legacy systems, to sourcing business operations.

About CognizantCognizant (NASDAQ: CTSH) is a leading provider of information technology, consulting, and business process services, dedicated to helping the world’s leading companies build stronger busi-nesses. Headquartered in Teaneck, New Jersey (U.S.), Cognizant combines a passion for client satisfaction, technology innova-tion, deep industry and business process expertise, and a global, collaborative workforce that embodies the future of work. With over 100 development and delivery centers worldwide and ap-proximately 233,000 employees as of March 31, 2016, Cogni-zant is a member of the NASDAQ-100, the S&P 500, the Forbes Global 2000, and the Fortune 500 and is ranked among the top performing and fastest growing companies in the world. Visit us online at www.cognizant.com or follow us on Twitter: Cognizant.